Optical communications systems can suffer from many problems and phenomenon, particularly systems which carry a large number of high capacity channels. One particular problem is non-linear optical interactions between signals. This problem is more severe in systems using low dispersion fibers, such as True Wave Classic, True Wave RS, and LEAF, but non-linear interactions can also be problematic in systems which do not use low dispersion fibers. Non-linear effects can include cross-phase modulation, self-phase modulation, and four wave mixing. The penalties associated with non-linear interactions can adversely affect the performance of a system, such as by reducing the signal population which can propagate the desired distance on a given fiber type, and by reducing the reach of the system.
Several prior art solutions have been proposed to address the problems of non-linear interactions. Unfortunately, all of the prior art solutions have significant disadvantages. One prior art solution is to operate in the “L-band” rather than the “C-band”. Although that solution would solve some of the problems caused by non-linear interactions, it has several serious drawbacks. For example, L-band components are not as readily available as C-band components, resulting in substantially increased costs. Furthermore, a substantial engineering investment will be required to produce an L-band product suite. As a result, this solution will be very expensive to implement, both in monetary cost and development time.
Another prior art solution is to reduce the number of channels in the C-band, and/or to reduce the launch power and reach of the system. While these solutions can reduce the effects of non-linear interactions by spacing signal channels farther apart and by eliminating poorly performing channels, it demands that the system performance (e.g., reach and/or capacity) be reduced. Although the system performance can be increased through the use of additional regenerators, regenerators are expensive and add to the cost of the system.
FIG. 1 illustrates an exemplary prior art dispersion map using True Wave Classic fiber. In that example, there are seven links, each link with five spans, and each span being eighty kilometers long. In that example, negative dispersion compensation is provided in each transmit node using a commercially available dispersion compensation module. The negative dispersion compensation is used to offset the positive dispersion introduced by the fibers over which the signals travel. The dispersion compensation can be chosen so that the center band (1544.9 nm), or another portion of the band, is well compensated. However, other portions of the band, such as the signals at edges (e.g., 1530.3 nm and 1563.5 nm), experience significant residual dispersion (e.g., about −750 ps/nm and +500 ps/nm, respectively, after propagating 2800 km). Further, the low wavelength portion of the C-band also experiences high non-linear penalties on this dispersion map. As a result, this prior art technique provides, at best, a partial solution to the problem, and introduces significant other problems in the form of the high dispersion variations at the edge of the band. Note that other prior art dispersion maps, which distribute the total dispersion compensating fiber required into smaller units in both the nodes and the spans, have even greater nonlinear penalties.
Another prior art solution is to individually dispersion compensate optical signal channels. While that solution may succeed in overcoming problems of high dispersion variation described above, it requires significant equipment and a significant increase in cost. In many systems, where the number of channels is very high, this solution cannot be used.
Another problem with the prior art is that standard dispersion compensating fiber is poorly slope-matched to some types of fiber, such as True Wave Classic. As a result, even for perfectly compensated links, signals at the edges of the C-band will encounter substantial residual dispersion at the end of long links. Although the length of the link will vary depending on many factors in the link, a typical link of 1,000 km will likely experience substantial residual dispersion. The penalty can be so high that some systems will not achieve acceptable bit-error-rate performance. The problem is further exacerbated by the fact that it is not feasible to perfectly compensate installed links, given the variation in the dispersion compensation fiber (DCF) coils and variation in length and dispersion of spans.
FIG. 2 is a diagram of OSNR Penalty verses residual dispersion illustrating the impact of poor slope compensation as may be seen in the prior art. Residual dispersion at high and low ends of the C-band is so high (or low) that significant OSNR penalties are incurred. This figure illustrates an example using True Wave Classic, compensated every three spans at −6 dBm.
One prior art solution is to use dispersion tolerant modulation modalities. This limits the user to relatively low bit rates or complex modalities. A second solution is to install dispersion compensation for blocks of wavelengths. Whether multiple dispersion compensation paths are installed along the length of the link or whether blocks of wavelengths are specially compensated at the end of the link, this adds cost to the network.
Accordingly, there is a need for systems, devices, and methods to control or compensate for non-linear interactions and which also provides for good system performance. Those and other advantages of the present invention will be described in more detail hereinbelow.